Adaptive Magnetic Resonance

Abstract

Nuclear magnetic resonance is one of the cornerstones of modern medicine and biomedical research. Over the past several decades, the speed and precision of in-vivo magnetic resonance imaging (MRI) and spectroscopy (MRS) have increased by leaps and bounds, by utilizing sophisticated excitation and acquisition techniques, from parallel imaging and compressed sensing to magnetic resonance fingerprinting. However, these approaches have all been static in nature, fixing measurement parameters in advance, in anticipation of a wide range of expected tissue parameter values, and are therefore sub-optimal for any given subject. We depart from the conventional framework of magnetic resonance and propose a new approach - termed adaptive magnetic resonance - which binds acquisition and excitation, by using the measured signal to update and fine-tune the measurement parameters in real time. This targets the specific tissue characteristics of the subject while they are being scanned. Adaptive magnetic resonance provides a completely new and previously-untapped avenue for improving the sensitivity and specificity of in-vivo magnetic resonance across all tissue contrast mechanisms. Equivalently, it can accelerate data acquisition compared to non-adaptive schemes, by obtaining the same precision using fewer, optimally tuned excitations. We demonstrate that an adaptive pulse sequence for measuring the transverse relaxation time (T2) of metabolites in-vivo improves upon the precision of static approaches by a factor of {approx} 1.7 - or, alternatively, accelerates acquisition 2.5-fold.